tools_schema

Instructions

Implementation Reference

• src/index.js:46-46 (registration)

  The tool 'tools_schema' is registered in the STATIC_TOOLS array as a static tool catalog entry used for inspection/fallback. It has an inputSchema requiring a 'tool_name' string property.

  { name: "tools_schema", description: "Get JSON schema for a tool — column_mapping and module_parameters required before tools_run.", inputSchema: { type: "object", properties: { tool_name: { type: "string", description: "Name of the tool" } }, required: ["tool_name"] } },

• src/index.js:121-146 (handler)

  The CallToolRequestSchema handler proxies all tool call requests (including tools_schema) to the remote MCP server via remoteClient.callTool().

  server.setRequestHandler(CallToolRequestSchema, async (request) => {
    if (!remoteClient) {
      return {
        content: [
          {
            type: "text",
            text: "MCP Analytics API key required. Set MCP_ANALYTICS_API_KEY in your environment.\nGet a free key at https://app.mcpanalytics.ai",
          },
        ],
        isError: true,
      };
    }
  
    try {
      const result = await remoteClient.callTool({
        name: request.params.name,
        arguments: request.params.arguments || {},
      });
      return result;
    } catch (err) {
      return {
        content: [{ type: "text", text: `Error: ${err.message}` }],
        isError: true,
      };
    }
  });

• src/index.js:70-103 (helper)

  When an API key is present, a remote client connects to the MCP API and fetches the live tool catalog (which includes tools_schema handler logic from the server side).

  if (API_KEY) {
    try {
      const transport = new StreamableHTTPClientTransport(
        new URL(`${API_URL}/mcp/api-key`),
        { requestInit: { headers: { "X-API-Key": API_KEY } } }
      );
  
      remoteClient = new Client({
        name: "mcp-analytics-proxy",
        version: "1.0.0",
      });
  
      await remoteClient.connect(transport);
  
      // Fetch live tool catalog
      const liveTools = [];
      let cursor;
      do {
        const { tools, nextCursor } = await remoteClient.listTools({ cursor });
        liveTools.push(...tools);
        cursor = nextCursor;
      } while (cursor);
  
      toolCatalog = liveTools;
      process.stderr.write(
        `[mcp-analytics] Connected to ${API_URL}. ${toolCatalog.length} tools available.\n`
      );
    } catch (err) {
      process.stderr.write(
        `[mcp-analytics] Could not connect to remote (${err.message}). Using static catalog (${STATIC_TOOLS.length} tools).\n`
      );
      remoteClient = null;
    }
  } else {

• src/index.js:46-46 (schema)

  Input schema for tools_schema: requires a 'tool_name' (string) property. The tool returns the JSON schema for the specified analysis tool's column_mapping and module_parameters.

  { name: "tools_schema", description: "Get JSON schema for a tool — column_mapping and module_parameters required before tools_run.", inputSchema: { type: "object", properties: { tool_name: { type: "string", description: "Name of the tool" } }, required: ["tool_name"] } },